Before signing a wallet request, the question is what permission it grants—not simply whether the website looks familiar. In a ChainCatcher report published November 11, 2026, Bitget Wallet said it would expand its AI-assisted threat detection, including checks intended to reveal hidden permissions before users sign.
What the increase measures
Bitget Wallet said the weekly average number of signature requests it flagged as potential phishing rose 7.4-fold in the third quarter of 2026: from about 5,000 in the second quarter to about 37,000. Chief operating officer Alvin Kan said the rise could reflect both more threat activity and improved detection, but the available data cannot measure each factor’s contribution. He stressed that the figures count flagged signing interactions, not distinct attacks or confirmed financial losses.
Bitget Wallet introduced AI-assisted risk detection in the second quarter. It said its expanded checks examine website code and domain registration information for signs of fraud, and simulate transactions before signing to look for hidden permissions. In October, AI identified 90% of newly blocked high-risk domains, compared with 36% in the second quarter. Nearly 70% of the domains flagged by AI that month had no matching alert in the external threat-intelligence sources Bitget Wallet checked.
The risks extend beyond warnings in one wallet. ChainCatcher reported that a trader lost nearly $1 million in July after signing a malicious Ethereum token approval. It also cited an August Rapid7 investigation that found a phishing campaign targeting 885,000 phone numbers and directing recipients to sites impersonating crypto wallet providers. These incidents are separate from Bitget Wallet’s count of flagged signing requests.
